The evolution of function is challenging to study, because it requires reconstructing reasonable models for extinct ancestral nodes. We propose to generate experimentally testable models for studying how evolution has introduced and modified functional relationships at the protein level associated with increased fitness. We complement the established statistical inference from sequence phylogenies (ancestral gene resurrection) with an analogous, but more radical procedure based on identifying common, core tertiary structures to reconstruct gene structure and function of enzymes far more ancient (albeit less secure) than those accessible from phylogenetic sequence-based methods. We focus on very ancient models for ancestral aminoacyl-tRNA synthetases, whose evolutionary descent was key to the origins of codon-directed protein synthesis and hence gene expression. The aaRS are not all homologous, but instead occur in two distinct superfamilies. This project is most deeply motivated by a desire to understand the profound symmetries that relate the two superfamilies. Among several hypotheses we hope to test is that the ancestral forms of class I and class II AARS were initially encoded on opposite strands of the same sense/antisense open reading frame. We introduce the term Urzymology (from Ur = primitive, original, early + enzyme) to describe the creation and experimental study of such ancestral proteins, which lie beyond the reach of ancestral gene resurrection. Urzymology brings with it the ability to manipulate biological objects across time. Complementation between Urzymes and subsequently acquired functional modules and parallel mutagenesis of Urzymes and contemporary enzymes make it possible to test explicit models for the evolution of catalysis, specificity, and allostery. Published proofs-of-principle for many obvious contingencies provide an exceptionally strong combination of transformative research. Aim 1 will document the relative amino acid specificities of Class I and II aminoacyl-tRNA synthetase Urzymes, and establish detailed mechanistic differences between the Urzymes and contemporary aaRS. Aim 2 is devoted to experimental study of the Rodin-Ohno hypothesis that the two aaRS classes arose on opposite strands of the same ancestral gene. Aim 3 will enhance the computational design process and establish genetic systems to select and characterize less cytotoxic constructs for eventual use in selecting Urzymes with improved enzymatic function. Charting the record of functional adaptation with experiments like those proposed here will complement the growing genomic sequence database by providing experimental tools to access and characterize likely evolutionary intermediates. Outlining the evolutionary record of functional adaptation will supplement intuitive use of sequence databases with experimental paradigms that complement drug design and the engineering and design of new protein reagents by explicit new understanding of how modules interact in proteins. Validating sense/antisense genetic coding would enrich understanding of the proteome, by identifying pairs of protein superfamilies that arose simultaneously, enhancing the meaning of homology.